A framework for multi-sensor integration and data acquisition

This thesis investigates and implements a framework for multi-sensor integration and data acquisition operating in the context of a robot workstation equipped with both contact and noncontact sensors. An appropriate low-level representation of raw data supplied by various sensors is developed. We propose the spacial proximity graph as a means to describe spatial relations between features of three-dimensional objects. We also examine two high-level modeling techniques, feature models, and the Hough shape transform.